Abstract
Background:
Daily walking paths exhibit varying environment features and require continuous adjustments to locomotor trajectories. Humans maintain lateral balance while navigating paths by modifying stepping in accordance with changing side-to-side path limitations (i.e. path width, lateral location). These processes are influenced by one’s actual physical ability to maintain balance, as well as their self-perceived balance ability. Older adults experience decreases in each of these abilities, which may alter their capacities to execute appropriate lateral stepping adaptations.
Research Question:
How do age, physical and self-perceived balance abilities interact to influence lateral stepping adaptations of older adults walking in complex environments with dynamic lateral path features?
Methods:
Twenty young (age mean±SD: 21.7±2.6) and 18 older adults (age mean±SD: 71.6±6.0) walked on an instrumented treadmill in a virtual-reality system. Participants adjusted lateral stepping during two competing lateral balance sub-tasks that manipulated either path width or location. Participants began walking on a gradually-narrowing path (sub-task A), then decided when/ how to exchange sub-tasks by laterally maneuvering to an adjacent path (sub-task B). Recorded path characteristics were used to quantify spatial thresholds for stepping error onset and sub-task exchange.
Results:
Older adults made sub-task A stepping errors on wider paths and exchanged sub-tasks earlier. These differences were not directly attributed to age. Statistical path analyses revealed that physical balance ability mediated age effects on stepping error onset, while self-perceived balance ability mediated age effects on sub-task exchange. Both age groups exchanged sub-tasks when stepping accuracy likelihoods were similar and high, ~90%.
Significance:
This work demonstrates important mechanisms for how age, via degradation of physical and self-perceived balance abilities, indirectly and differentially influences navigation of competing lateral balance tasks. Mediating physical and perceptual factors are potential targets for improving older adults’ navigation of complex environments.
Keywords: Walking, Lateral Balance, Aging, Perception, Lateral Maneuvers, Narrow-Base Walking
INTRODUCTION
Older adult falls frequently result from inadequate stepping responses to balance challenges encountered while walking [1, 2]. Dynamic properties of routine walking environments challenge balance by placing ever-changing spatial limitations on feasible walking trajectories. Real-world circumstances like navigating crowded sidewalks or store aisles thus require humans to seek out available walking paths around and between changing environment features [3]. Humans do this by selecting, then executing modified stepping strategies amidst infinite possibilities. Older adults may demonstrate important differences in these processes.
Humans continuously adjust lateral (i.e. side-to-side) stepping based on concurrent lateral path considerations, including both path width and lateral location of the path center [4]. Ecologically, both of these path features vary frequently and independently. Lateral stepping adaptations are often assessed by empirically manipulating either path width [5] or lateral location [6], but not both. Experimental paradigms could be improved by introducing concurrent manipulations of path width and location, such that path-navigation task complexity better approximates that encountered ecologically. In these, people must weigh the urgency of demands imposed by multiple path features, and the relative risks of different potential stepping strategies [7].
Specific environment features inform appropriate stepping adaptations [8]. However, a person’s physical ability to execute an intended adjustment in a given context is also paramount [9]. Connections are established between typical aging and physical balance limitations [10–12]. Separate connections are established between physical balance deficits and altered lateral stepping strategies [13, 14]. However, mechanisms detailing the extent to which aging might indirectly impede lateral stepping, via physical degradation, are not well defined. Physical deficits largely explain age-related changes in sagittal-plane stepping adjustments [15], but connections remain unknown in the frontal-plane.
A person’s self-perceived balance ability also impacts their interaction with challenging walking paths [16, 17]. Nearly 1/3 of older adults misjudge their balance abilities [18]. The extent of this misjudgment is characterized by the discrepancy between physical and self-perceived balance abilities. Thus, age-related changes in self-perceptions may further, or alternatively, explain altered elderly stepping adaptations [19]. Older adults’ under- or overestimation of their actual balance ability promote either overly-conservative, avoidance-based strategies [20], or undue risk-taking strategies leading to falls [16]. Again, the specific mechanisms whereby aging may indirectly influence lateral stepping via altered self-perception, have not been defined.
Here, we sought to clarify the mechanisms whereby age, physical balance ability, and self-perceived balance ability contribute as independent or interacting factors to alter older adults’ lateral stepping adaptations to concurrent lateral balance challenges. We introduced a complex, competing-tasks paradigm where young and older adults navigated two balance sub-tasks defined by varied path width or lateral location features. Participants walked on a constantly narrowing-path (path width manipulated) while explicitly deciding when/ how to maneuver onto an adjacent path (path lateral location manipulated). We determined the extents to which physical and self-perceived balance abilities explained age group differences in spatial thresholds of stepping error onset and sub-task exchange.
We predicted that older adults would exhibit earlier stepping error onset (at wider spatial thresholds) while traversing the constantly narrowing-path, as older adults demonstrate wider steps [5] and reduced stepping accuracy [17] on narrow paths. We further predicted that older adults would execute earlier path switches (at wider spatial thresholds), offsetting narrowing-path balance risks with more-conservative stepping strategies [21]. Statistical path models relating age to each of these two spatial thresholds via physical ability/ self-perception mediators likely vary differentially. We lastly defined a third relevant threshold which quantified (as a percentage) participants’ likelihood of stepping within the narrowing-path bounds when they switched paths. We expected that this likelihood threshold would either be much higher or lower in older adults, indicating relative risk-avoidance/acceptance adaptation strategies. With defined likelihood thresholds and mechanisms relating age to task-performance spatial thresholds, we may better identify and target the specific ways in which age and related factors alter stepping adaptations on challenging walking paths.
METHODS
Participants
Twenty young healthy (YH) and 20 older healthy (OH) adults provided written, informed consent prior to participation. This study was approved by The Pennsylvania State University’s IRB. Participants were screened to ensure that they had no cardiovascular, neurological, visual, or musculoskeletal conditions which might have affected participation. Participants were required to score ≥ 24/30 on the Mini-Mental State Exam [22]. Two OH participants relied heavily on handrails during treadmill walking trials. Their data were excluded from all analyses. Two-sample t-tests or Mann-Whitney U tests (for data not normally distributed) assessed age group differences in participant characteristics and assessment scores (Table 1).
Table 1 –
Young Healthy and Older Healthy general physical characteristics were collected. Functional lower extremity strength, functional mobility and dynamic balance were assessed using 30-second Chair Stand Test (30CST), Timed Up and Go (TUG), and Four Square Step Test (FSST), respectively. Balance confidence and falling concern were assessed using the Activities-Specific Balance Confidence (ABC) Scale and Iconographical Falls Efficacy Scale (I-FES), respectively. Physical Ability composite scores were defined using the shown linear combination of standardized 30CST, TUG and FSST scores, which explained the most between-subject variance. Perceived Ability scores were similarly defined using the shown linear combination of standardized ABC and I-FES scores. All values except Sex are given as Mean ± Standard Deviation. Two-sample t-test or Mann-Whitney U test p-values indicating significant group mean differences are bolded.
Characteristic: | Young Healthy (YH): | Older Healthy (OH): | p-value |
---|---|---|---|
Sex [M/F] | 9 / 11 | 5 / 13 | |
Age [yrs] | 21.7 ± 2.6 | 71.6 ± 6.0 | < 0.001 |
Body Height [m] | 1.73 ± 0.08 | 1.70 ± 0.10 | 0.305 |
Body Mass [kg] | 69.9 ± 12.1 | 70.0 ± 10.5 | 0.983 |
Body Mass Index | 23.2 ± 2.7 | 24.2 ± 3.3 | 0.260 |
Leg Length [m] | 0.82 ± 0.06 | 0.84 ± 0.03 | 0.366 |
Assessment: | |||
30CST [stands/30s] | 17.8 ± 6.0 | 13.7 ± 3.32 | 0.002 |
TUG [s] | 6.62 ± 0.75 | 8.20 ± 1.92 | 0.004 |
FSST [s] | 7.63 ± 1.37 | 9.19 ± 1.53 | 0.002 |
ABC Score [/100 %] | 96.8 ± 2.5 | 93.1 ± 5.5 | 0.023 |
I-FES [/40] | 12.3 ± 1.9 | 16.1 ± 3.9 | 0.001 |
Composite Score: | |||
Physical Ability = −0.496 * (30CST) + 0.636 * (TUG) + 0.592 * (FSST) | 0.1 ± 3.2 | −3.8 ± 3.2 | <0.001 |
Perceived Ability = −0.707 * (ABC) + 0.707 * (I-FES) | 2.3 ± 1.5 | −0.8 ± 3.2 | <0.001 |
Physical Balance Abilities
We assessed participants’ overall performance on established functional assessments of several critical physical determinants of balance. We evaluated participants’ functional lower extremity strength (30-second Chair Stand Test; 30CST) [10], functional mobility (Timed Up and Go; TUG)[23], and dynamic balance (Four Square Step Test; FSST)[12]. For each subject, a composite “Physical Ability” score (Table 1) was calculated from their standardized 30CST, TUG and FSST scores using principal components analysis [15]. Each composite score was the linear combination of assessment z-scores that explained the most data variance [24].
Self-Perceived Balance Abilities
We assessed participants’ overall responses on established questionnaires assessing self-perceptions of balance ability. We evaluated participants’ balance confidence (Activities-Specific Balance Confidence Scale (ABC)) [25] and falling concern (abbreviated Iconographical-Falls Efficacy Scale (I-FES)) [26]. For each participant, a composite “Perceived Ability” score (Table 1) was calculated from their standardized ABC and I-FES scores using principal components analysis [15].
Walking Sub-Tasks
Participants walked in an “M-gait” virtual reality system (Fig. 1A–B). For all walking trials, treadmill speed was fixed at 0.75 m/s (optic flow matched) to accommodate potential preference for reduced walking speeds while narrow-path walking [5, 21], without introducing confounding effects of self-selected walking speeds. First, participants acclimated with four minutes of treadmill walking. Several subsequent warm-up trials progressively introduced elements of path walking, path switching, and stepping error feedback. Participants then completed five, 4-minute path-navigation trials. Each trial consisted of eight bouts involving two balance sub-tasks. For each bout, participants began walking on a continuously narrowing-path (sub-task A; path width consideration) then chose when to switch to an adjacent path (sub-task B; path location consideration) (Fig. 1). Thus, participants adapted stepping to sub-task A or sub-task B path-defined spatial limitations. Participants were encouraged to minimize stepping errors outside of any path bounds. Auditory feedback was provided. Narrowing-paths on the right vs. left side of the treadmill were alternated. Participants rested (≥ 2-min) and performed a normal-walking ‘wash-out’ trial (60s) between path-navigation trials.
Figure 1 –
Participants walked in an “M-gait” virtual reality system, which integrates an instrumented treadmill, virtual reality, and motion capture (Motekforce Link, Amsterdam, Netherlands). Participants wore a ceiling-mounted safety harness and were encouraged to use handrails only to avoid falling. In each trial, participants traversed a series of 25m-long walking bouts that included two sub-tasks, defined by systematically-varied lateral path features (path width or lateral location). For sub-task A, participants began each bout on a virtual path that gradually narrowed from 0.45m to 0.05m at a rate of 0.012m/s (A). Participants adjusted lateral stepping to accurately walk within the narrowing-path bounds and received auditory feedback of any stepping errors. Meanwhile, an adjacent path remained at a fixed width of 0.45m. For balance sub-task B, participants maneuvered from the narrowing-path to the adjacent, wider path (B). Participants decided when and how to execute sub-task B. Sub-task A risk and difficulty increased as the path narrowed. Concurrently, sub-task B risk and difficulty also increased, as the narrowing-path increased the distance between adjacent path bounds. Thus, participants were required to make an explicit decision to trade-off the sub-task balance challenges imposed by the continuously-changing path width vs. path location demands. CoM trajectories (orange) and first stepping errors (red foot) were recorded to distinguish various spatial thresholds defined by the width of the narrowing-path at key instances of stepping error onset (werror1) and sub-task exchange (wswitch) (C).
Data Collection and Analyses
Each subject wore 16 retroreflective markers. Four markers each were placed on the head (left/ right fore/ backhead), pelvis (left/ right Posterior Superior Iliac Spine (PSIS) and Anterior Superior Iliac Spine (ASIS)), and on the shoe surfaces, aligned with foot landmarks (lateral malleolus, calcaneus, first and fifth metatarsal heads). Raw data were collected at 120 Hz and processed using Vicon Nexus (Oxford Metrics, Oxford, UK) and D-Flow (Motekforce Link, Amsterdam, Netherlands). Marker trajectories were analyzed in MATLAB (MathWorks, Natick, MA).
We analyzed 35 instances (latter 7 of 8 bouts for each of 5 trials) of narrowing-path walking/ switches for each participant. Marker trajectory data were first low-pass filtered (4th-order Butterworth, cutoff: 10 Hz) and interpolated (600 Hz). Each step was defined at heel strike [27]. Each path-switch occurred when the pelvic centroid crossed the treadmill midline (Fig. 1C). The width of the narrowing-path was extracted at stepping and switching events.
Sub-Task Spatial and Likelihood Thresholds
For each bout, narrowing-path width at first stepping error, werror1, quantified the spatial threshold for sub-task A stepping error onset. The narrowing-path width at the instant of path switching, wswitch, quantified the spatial threshold at sub-task B execution (thereby at sub-task exchange). These spatial thresholds were computed for all path walking/ switching instances, then averaged for each participant.
We then assessed stepping accuracy likelihood across narrowing-path widths. Similarly-quantified likelihoods have provided insights regarding older adult risk-aversion [16] and balance ability misjudgment [17], but only using paths with static features. Our narrowing-path paradigm facilitated approximations of group-wise stepping accuracy dynamics as continuous functions of narrowing-path width. For each step during sub-task A, the leading foot’s mediolateral position and current narrowing-path width were determined, then compiled across subjects into YH and OH groups (Fig. 3A–B). Foot placement data were distributed into ‘sliding’ bins according to corresponding narrowing-path widths, then sub-sampled. Each bin’s sub-samples corresponded with one quasi-static path width, the mid-point of that bin’s range. Gaussian probability curves were fit to each sub-sample of foot position data (Fig. 3C–D). Pin quantified the percentage of a given Gaussian distribution area contained within quasi-static path boundaries, thereby representing the percentage likelihood of stepping within those bounds [16, 17]. For each group, logistic curves of the form Pin = 100% / (1+Ae−b[Path Width]) were fit to Pin vs. Path Width to characterize stepping accuracy dynamics and the group-wise Pin rate of change (b) on the narrowing-path. We mapped each participant’s wswitch onto these continuous Pin curves at Pswitch. This Pswitch defined each participant’s stepping accuracy likelihood threshold at sub-task exchange. Higher Pswitch indicate greater risk-avoidance and lower sub-task A stepping error tolerance.
Figure 3 –
Mediolateral foot centroids (relative to path center) and corresponding narrowing-path widths at each step throughout sub-task A were determined for each bout, then compiled across subjects into YH and OH groups (A-B). The last step prior to each transition was considered part of the transition strategy and was removed from compiled data. For each group, all foot positions were placed into ‘sliding’ bins according to corresponding narrowing-path widths between 0.05 and 0.45 m. Each bin ranged 5 mm of path width and moved every 2.5 mm. Bins containing fewer than 20 steps were removed from the analysis. For each bin, 50 data sets were sub-sampled based on a random 95% of foot position data in that bin. Each bin’s sub-samples corresponded with one quasi-static path width, the mid-point of that bin’s range. For each of the sub-sampled data sets, a Gaussian probability curve was fitted to the foot position data. These Gaussian distributions are plotted at the corresponding bin’s quasi-static path width (C-D). Walking path bounds are shown with colored tiles in the ML Foot Placement vs. Path Width plane. The percentage of each distribution’s area contained within the path bounds represents Pin, the group-wise likelihood of a participant stepping within the path bounds at the given path width. Pin decreases as sub-task A path width decreases, seen qualitatively as distributions for smaller path widths are not fully contained by the path bounds.
Contributions of Age, Physical Ability and Self-Perceived Ability
All statistical analyses were performed in Minitab (Minitab, Inc., State College, PA). Mann-Whitney U tests compared age group differences in spatial thresholds werror1 and wswitch. Pairwise Pearson correlations related these thresholds with Physical Ability and Perceived Ability composites. Based on these correlations, apparent age group differences in werror1 and wswitch may be confounded. To test this, age group comparisons were repeated by producing generalized linear models with Physical Ability and Perceived Ability scores as covariates. These tests indicated whether physical and self-perceived ability confound spatial threshold age differences, but not to what extent.
Multiple regression analyses determined the relative contributions of age, physical ability and self-perceived ability factors to each spatial threshold. We first considered that each factor may directly influence werror1 and wswitch. Alternatively, age effects may be mediated by physical and/ or self-perceived balance ability. Statistical path analyses tested these mechanisms, assessing physical ability and perceived ability as either independent contributors or mediators of age on werror1 and wswitch [24]. We also considered a potential interaction between physical and self-perceived ability factors. All analyses used standardized z-scores. Pathways with non-significant coefficients (p>0.05) were eliminated in final path models. A χ2-test evaluated goodness-of-fit for each final model.
RESULTS
Age groups did not significantly differ in measured characteristics (Table 1). However, OH participants demonstrated reduced functional lower extremity strength, functional mobility, and dynamic stepping ability compared to YH. These were indicated by lower 30CST, and longer TUG and FSST times, respectively (Table 1). Consequently, OH composite Physical Ability scores were significantly lower, representing a reduced overall physical balance ability.
OH participants demonstrated lower balance confidence and greater falling concern than YH. These were indicated by lower ABC scores and higher I-FES, respectively (Table 1). Consequently, OH composite Perceived Ability scores were significantly lower, representing a reduced overall self-perceived balance ability.
YH and OH similarly executed sub-task B after 1–2 sub-task A stepping errors (YH: 1.5±0.7; OH: 1.7±1.0; p=0.589). OH exhibited stepping errors sooner, at wider werror1 (Fig 2A). OH exchanged sub-tasks sooner, at wider wswitch (Fig. 2B).
Figure 2 –
Older adults demonstrated wider spatial thresholds for sub-task A stepping error onset (werror1), making initial stepping errors sooner, at wider narrowing-path widths (YH: 0.138 ± 0.039 m; OH: 0.202 ± 0.075 m; p = 0.003) (A). Older adults also demonstrated wider sub-task exchange spatial thresholds (wswitch), as they more-readily executed sub-task B, switching walking paths sooner, at wider narrowing-path widths (YH: 0.113 ± 0.025 m; OH: 0.161 ± 0.070 m; p = 0.022) (B). Both spatial thresholds were defined by narrowing-path widths at error and path-switching instances and were indirectly related to sub-task A difficulty and risk. Asterisks indicate significant group differences.
Qualitatively, OH foot placements were wider and more variable (Fig. 3A–B). For both groups, stepping accuracy likelihood (Pin) decreased as the sub-task A path narrowed (Fig. 3C–D). Continuous, group-wise curves of accuracy likelihood against narrowing-path width (Fig. 4A–B) demonstrated that YH Pin diminished later, at narrower path widths, but degraded more rapidly (−b = −0.67 m−1). OH Pin diminished sooner, but did so more gradually (−b = −0.33 m−1). Likelihood threshold Pswitch group means were above 90% and did not significantly differ (Fig. 4C). Thus, while OH exchanged sub-tasks at wider spatial thresholds than YH, both groups exchanged tasks when likelihood thresholds were similar.
Figure 4–
For each of the ‘sliding’ bins from Fig. 3, Pin quantifies the percent likelihood of stepping within a given bin’s quasi-static path boundaries for a given path width. These probabilities for each distribution (from 50 sub-sampled data sets) were averaged for each bin and plot as Pin vs. Path width for YH (A; blue) and OH (B; red). A logistic curve of the form Pin = 100% / (1+Ae−b[Path width]) was fit to each set of probability data to determine the group rate at which sub-task A stepping accuracy changed with path width (b). Then, each participant’s wswitch was mapped onto the respective group curves to determine participants’ stepping accuracy likelihood threshold Pswitch, at sub-task exchange (instance demonstrated in B). Pswitch did not significantly differ between groups (YH: 91.2 ± 6.9 %; OH: 92.9 ± 6.8 %; p = 0.373) (C). Groups thus had similar likelihoods of sub-task A stepping accuracy when they exchanged sub-tasks, despite doing so at significantly different path widths (Fig. 2B).
Both werror1 and wswitch were negatively correlated with composite scores (Fig. 5A–D). Age effects for werror1 were no longer significant once Physical Ability composites, Perceived Ability composites, then both were included as covariates (p=0.083, p=0.051 and p=0.222). Group effects for wswitch were also no longer significant (p=0.131, p=0.175, p=0.371). Thus, group differences in both spatial thresholds were explained by physical and self-perceived ability differences, independent of differences in age per se.
Figure 5 –
Sub-task spatial thresholds werror1 and wswitch were significantly negatively correlated with each Physical Ability and Perceived Ability composites via Pairwise Pearson correlations (A-D). Thus, individuals who demonstrated greater balance-relevant physical ability and perceived ability tended to maintain sub-task A success longer by avoiding stepping error onset (narrower werror1), and also tended to exchange sub-tasks later (narrower wswitch). Differences in Physical and Perceived Ability composite scores largely explained the age group differences in spatial thresholds. The most-likely statistical path models were those in which Physical Ability composites significantly mediated the effect of Age on werror1 (E), while Perceived Ability composites significantly mediated the effect of Age on wswitch (F). Each statistical path model presented in E–F consists of two predictive relations, represented by separate regression equations. Each equation relates i and j factors, where cij are the regression coefficients and εij are the error terms. χ2 tests for final path models of werror1 (χ2 = 1.26; p = 0.53) and wswitch (χ2 = 1.06; p = 0.59) indicated agreement between the data and statistical path models (p >> 0.05).
Statistical path analyses confirmed that the least likely statistical path models (no significant regression coefficients) were those in which combined Age, Physical Ability and Perceived Ability independently influenced werror1 and wswitch. Age was a significant predictor of both Physical Ability and Perceived Ability (p=0.001; p=0.001). Physical Ability significantly influenced werror1 (p=0.001), while Perceived Ability did not (p=0.090) (Fig. 5E). Perceived Ability significantly influenced wswitch (p=0.001), while Physical Ability did not (p=0.120) (Fig. 5F). Final statistical path models demonstrate that Age indirectly influenced werror1 and wswitch via Physical Ability and Perceived Ability mediators, respectively. Physical and Perceived Ability scores were significantly correlated (r=0.538, p<0.001), but did not interact in the final statistical path models.
DISCUSSION
In complex environments, humans maintain balance by adjusting foot placements [28] to accommodate environment-defined spatial limitations [9]. Altered lateral stepping strategies are sensitive to specific balance demands [29], frequently imposed by some combination of constrained path width and/or lateral location [4]. Empirical manipulations of just path width or location may over-simplify the balance task complexity required for real-world path navigation. Consequently, these may not highlight underlying deficits impacting balance [30]. Here, we introduced a VR-based paradigm which presented complex walking paths with concurrent, systematically-varied lateral features. We characterized young and older adults’ spatial thresholds at instances of stepping error onset and sub-task exchange. Statistical path models detailed how aging, via degradation of physical and self-perceived balance abilities, impacted lateral stepping adjustments in task-specific ways. Thresholds of stepping accuracy likelihood reflected similar risk-taking, independent of age.
OH adults’ 30CST scores (Table 1) were lower than YH, but comparable to age-matched norms [10]. This OH cohort was still quite fit, as TUG and FSST times were slower than YH, but faster than age-matched norms [11, 12]. No scores assessing physical ability determinants surpassed assessment-defined thresholds for falling concern. Nonetheless, OH’s demonstrated physical balance deficits relative to YH for all determinants assessed. We have succinctly represented this in Physical Ability composite scores, which were lower for OH.
Similarly, OH adults demonstrated consistently lower ABC and higher I-FES relative to YH (Table 1), but scores were far from falling concern thresholds [26, 31]. Still, these indicated OH’s relatively lower balance confidence and greater fear of falling, indicating general self-perceptions as less-capable of maintaining balance. This is succinctly reflected in Perceived Ability composite scores, which were lower for OH.
As predicted, OH required wider paths to avoid stepping errors (Fig. 2A). This is consistent with findings that older adults demonstrate reduced stepping accuracy [17] and larger step widths [5] during narrow-base walking. This apparent age effect was explained with a statistical path model where physical ability mediated the age-effect on werror1 (Fig. 5E). This makes sense, as strength, mobility and dynamic balance likely impact active control of foot placement.
OH exchanged sub-tasks earlier, at wider paths than YH, as predicted (Fig. 2B). While this study did not directly assess participants’ perceptions of sub-task relative difficulties, wswitch indicates a stepping adaptation following explicit re-prioritization of balance demands. The wswitch group differences were dominantly mediated by participants’ self-perceived balance ability (Fig. 5F), meaning earlier OH path-switches were motivated by perceived inabilities to execute either sub-task. Our identification of exact wswitch thresholds on continuously narrowing-paths extends previous work relating older adults’ balance self-perceptions to their selections of fixed-width walking surfaces [16]. Older adults’ wider wswitch may have been further motivated by their earlier sub-task A error onset, as wswitch and werror1 were highly correlated across participants (r=0.90, p<0.001) and participants readily exchanged sub-tasks following 1–2 errors. Wider werror1 and wswitch indicated OH’s more-conservative stepping adaptations to several changing path limitations. Aging indirectly and differentially altered each spatial threshold, where mediators were different for error onset versus sub-task exchange thresholds.
Fig. 5E–F mechanisms do not consider other potential mediators not empirically-evaluated in this current work focusing on physical vs. self-perceived ability contributions. Older adults’ wider werror1 and wswitch could be further explained by cognitive mediators like age-related attentional deficits. These deficits may limit older adults’ concurrent task-performance, subsequently impacting either spatial threshold, but also tend to be confounded by physical balance capacity [30]. Alternatively, older adults’ wider spatial thresholds may arise from reduced stepping-precision capabilities, especially while narrow-base walking, caused by age-related increases in sensorimotor noise [32]. Identifying such additional mediating factors is an objective for future studies.
This work is the first demonstrating continuous functions predicting group-wise stepping accuracy degradation with decreasing path widths. Despite mediated age-differences in when age group exchanged sub-tasks, YH and OH did so when Pswitch were similar, ~90% (Fig. 4). Our results provide objective evidence for the relevance of a 90% likelihood threshold, previously defined arbitrarily for paradigms using several fixed-width paths to assess stepping accuracy [16, 17]. Likelihood thresholds provided bases for evaluating participants’ risk-taking, as higher percentage likelihoods indicate risk-avoidance strategies and lower likelihoods indicate risk-accepting strategies. Contrary to our predictions, OH demonstrated similarly-high likelihood thresholds as YH, indicating a comparable level of risk-avoidance. However, OH’s wider spatial thresholds indicate that, in order to similarly avoid risk, OH had to interact differently with the varied path features.
In this work, we defined distinct mechanisms demonstrating that older adults’ lateral stepping adaptations are not altered by age per se, but mediated by physical and self-perceived balance factors, in task-dependent ways. Physical balance ability mediated error-avoidance adaptations to narrowing-path widths, while perceived ability mediated sub-task exchange (path-switching) adaptations. OH retained similar and high stepping accuracy likelihoods by accepting wider spatial thresholds. Thus, priorities to avoid risk and preserve stepping accuracy are apparently preserved with healthy aging, but require older adults to alter how they interact with the spatial constraints of their walking environments. This work conveys the importance of considering how healthy aging influences distinct interactions between intrinsic, modifiable falling risk-factors including physical and self-perceived balance abilities, with extrinsic risk-factors including walking path limitations. Balance-training programs that address these interactions will better improve older adults’ stepping adaptability in challenging walking environments.
Highlights:
Spatial limitations of environment-defined walking paths create balance challenges
Humans navigate complex paths by adjusting lateral steps to multiple balance tasks
During competing tasks, older adults make stepping errors and exchange tasks sooner
Healthy aging preserves accuracy thresholds but alters how adults interact w/ paths
Physical & self-perceived balance deficits mediate age effects on lateral stepping
ACKNOWLEDGEMENTS
The authors would like to acknowledge Anna C. Render and David M. Desmet for their contributions and technical support throughout data collections.
This work was supported by NIH grant 1-R01-AG049735.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
CONFLICT OF INTEREST
The authors declare that there is no conflict of interest associated with this work.
References
- [1].Robinovitch SN, Feldman F, Yang Y, Schonnop R, Leung PM, Sarraf T, et al. , Video capture of the circumstances of falls in elderly people residing in long-term care: an observational study, Lancet 381 (9860) (2013) 47–54, 10.1016/S0140-6736(12)61263-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Kelsey JL, Procter-Gray E, Hannan MT, Li W, Heterogeneity of falls among older adults: implications for public health prevention, Am. J. Public Health 102 (11) (2012) 2149–2156, 10.2105/AJPH.2012.300677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Moussaïd M, Helbing D, Theraulaz G, How simple rules determine pedestrian behavior and crowd disasters, Proc. Natl. Acad. Sci. U. S. A 108 (17) (2011) 6884–6888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Dingwell JB, Cusumano JP, Humans use multi-objective control to regulate lateral foot placement when walking, PLoS Comput. Biol 15 (3) (2019), 10.1371/journal.pcbi.1006850. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Schrager MA, Kelly VE, Price R, Ferrucci L, Shumway-Cook A, The effects of age on medio-lateral stability during normal and narrow base walking, Gait Posture 28 (2008) 466–471, 10.1016/j.gaitpost.2008.02.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Hsieh KL, Sheehan RC, Wilken JM, Dingwell JB, Healthy individuals are more maneuverable when walking slower while navigating a virtual obstacle course, Gait Posture 61 (2018) 466–472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Braun DA, Nagengast AJ, Wolpert D, Risk-sensitivity in sensorimotor control, Front. Hum. Neurosci 5 (2011), 10.3389/fnhum.2011.00001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Twardzik E, Duchowny K, Gallagher A, Alexander N, Strasburg D, Colabianchi N, et al. , What features of the built environment matter most for mobility? Using wearable sensors to capture real-time outdoor environment demand on gait performance, Gait Posture 68 (2019) 437–442, 10.1016/j.gaitpost.2018.12.028. [DOI] [PubMed] [Google Scholar]
- [9].Fajen BR, Perceiving possibilities for action: on the necessity of calibration and perceptual learning for the visual guidance of action, Perception 34 (6) (2005) 717–740, 10.1068/p5405. [DOI] [PubMed] [Google Scholar]
- [10].Jones C, Rikli R, Beam W, A 30-s chair-stand test as a measure of lower body strength in community-residing older adults, Res. Q. Exerc. Sport 70 (2) (1999) 113–119, 10.1080/02701367.1999.10608028. [DOI] [PubMed] [Google Scholar]
- [11].Ibrahim A, Singh DA, Shahar S, ‘Timed up and go’ test: age, gender and cognitive impairment stratified normative values of older adults, PLoS One 12 (10) (2017), 10.1371/journal.pone.0185641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Dite W, Temple VA, A clinical test of stepping and change of direction to identify multiple falling older adults, Arch. Phys. Med. Rehabil 83 (11) (2002) 1566–1571, 10.1053/apmr.2002.35469. [DOI] [PubMed] [Google Scholar]
- [13].Shin S, Valentine J, Evans E, Sosnoff J, Lower extremity muscle quality and gait variability in older adults, Age Ageing 41 (2012) 595–599, 10.1093/ageing/afs032. [DOI] [PubMed] [Google Scholar]
- [14].Sawers A, Hafner B, Validation of the narrowing beam walking test in lower limb prosthesis users, Arch. Phys. Med. Rehabil 99 (8) (2018) 1491–1498, 10.1016/j.apmr.2018.03.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Kang HG, Dingwell JB, Separating the effects of age and speed on gait variability during treadmill walking, Gait Posture 27 (4) (2008) 572–577, 10.1016/j.gaitpost.2007.07.009. [DOI] [PubMed] [Google Scholar]
- [16].Butler AA, Lord SR, Taylor JL, Fitzpatrick RC, Ability versus hazard: risk- taking and falls in older people, J. Gerontol. A Biol. Sci. Med. Sci 70 (5) (2015) 628–634, 10.1093/gerona/glu201. [DOI] [PubMed] [Google Scholar]
- [17].Kluft N, van Dieën JH, Pijnappels M, The degree of misjudgement between perceived and actual gait ability in older adults, Gait Posture 51 (2017) 275–280, 10.1016/j.gaitpost.2016.10.019. [DOI] [PubMed] [Google Scholar]
- [18].Delbaere K, Close JCT, Brodaty H, Sachdev P, Lord SR, Determinants of disparities between perceived and physiological risk of falling among elderly people: cohort study, BMJ 341 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Maki BE, Gait changes in older adults: predictors of falls or indicators of fear? J. Am. Geriatr. Soc 45 (3) (1997) 313–320, 10.1111/j.1532-5415.1997.tb00946.x. [DOI] [PubMed] [Google Scholar]
- [20].Delbaere K, Crombez G, Vanderstraeten G, Willems T, Cambier D, Fear-related avoidance of activities, falls and physical frailty. A prospective community-based cohort study, Age Ageing 33 (4) (2004) 368–373, 10.1093/ageing/afh106. [DOI] [PubMed] [Google Scholar]
- [21].da Silva Costa A, Moraes R, Hortobágyi T, Sawers A, Older adults reduce the complexity and efficiency of neuromuscular control to preserve walking balance, Exp. Gerontol 140 (2020), 10.1016/j.exger.2020.111050. [DOI] [PubMed] [Google Scholar]
- [22].Folstein MF, Folstein SE, McHugh PR, “Mini-mental state”: a practical method for grading the cognitive state of patients for the clinician, J. Psychiatr. Res 12 (3) (1975) 189–198, 10.1016/0022-3956(75)90026-6. [DOI] [PubMed] [Google Scholar]
- [23].Podsiadlo D, Richardson S, The timed up and go test - a test of basic functional mobility for frail elderly persons, J. Am. Geriatr. Soc 39 (2) (1991) 142–148. [DOI] [PubMed] [Google Scholar]
- [24].Dingwell JB, Cavanagh PR, Increased variability of continuous overground walking in neuropathic patients is only indirectly related to sensory loss, Gait Posture 14 (1) (2001) 1–10, 10.1016/S0966-6362(01)00101-1. [DOI] [PubMed] [Google Scholar]
- [25].Powell LE, Myers AM, The activities-specific balance confidence (ABC) scale, J. Gerontol. A Biol. Sci. Med. Sci 50A (1) (1995) M28–M34, 10.1093/gerona/50A.1.M28. [DOI] [PubMed] [Google Scholar]
- [26].Delbaere K, Smith ST, Lord SR, Development and initial validation of the iconographical falls efficacy scale, J. Gerontol. A Biol. Sci. Med. Sci 66A (6) (2011) 674–680, 10.1093/gerona/glr019. [DOI] [PubMed] [Google Scholar]
- [27].Zeni JA, Richards JG, Higginson JS, Two simple methods for determining gait events during treadmill and overground walking using kinematic data, Gait Posture 27 (4) (2008) 710–714, 10.1016/j.gaitpost.2007.07.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Bruijn SM, van Dieën JH, Control of human gait stability through foot placement, J. R. Soc. Interface 15 (143) (2018) 1–11, 10.1098/rsif.2017.0816. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Kazanski ME, Dingwell JB, Cusumano JP, How healthy older adults regulate lateral foot placement while walking in laterally destabilizing environments, J. Biomech 104 (2020), 10.1016/j.jbiomech.2020.109714. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Woollacott M, Shumway-Cook A, Attention and the control of posture and gait: a review of an emerging area of research, Gait Posture 16 (1) (2002) 1–14, 10.1016/S0966-6362(01)00156-4. [DOI] [PubMed] [Google Scholar]
- [31].Lajoie Y, Gallagher SP, Predicting falls within the elderly community: comparison of postural sway, reaction time, the Berg balance scale and the Activities-specific Balance Confidence (ABC) scale for comparing fallers and non- fallers, Arch. Gerontol. Geriatr 38 (1) (2004) 11–26, 10.1016/S0167-4943(03)00082-7. [DOI] [PubMed] [Google Scholar]
- [32].Dean JC, Alexander NB, Kuo AD, The effect of lateral stabilization on walking in young and old adults, IEEE Trans. Biomed. Eng 54 (11) (2007) 1919–1926, 10.1109/TBME.2007.901031. [DOI] [PubMed] [Google Scholar]